224 Alian R. Robinson and Jurgen Sellschopp
rine. Autonomous underwater vehicles (AUV) will gain more importance even for
overt surveys in future when the maximum range of small affordable units
increases to operationally meaningful distances. For time series at fixed locations,
bottom moored systems are appropriate, deployed from aircraft, ship or AUV. Following a pre-programmed sequence, they release a surface float with an antenna
that transmits the most recent data via satellite to the laboratory (Tyce et al, 1998).
Moorings of this kind are required also for non-denied observations because traditional moored instruments would store data during the full deployment period.
Data are read out only after recovery and thus not available in time for ocean prediction.
Open military environmental investigations like any civili an ocean research
activity mostly rely on ships as the main platforms. The expenses for ship time are
easily compensated by the advantage of hands-on interaction with measurement
systems. an a ship it is easier to meet the requirements for instrumentation, which
need not be as complex and sophisticated as in remotely operated systems.
For the analysis ofthe physical ocean, the direct measurement ofthose quantities
is most valuable that are represented in the fields of numeric al models, namely
temperature, salinity and water velocity. Temperature and conductivity below the
sea surface require direct probing. The limitation to one-dimensional measurements when in situ probes are lowered from a ship, is overcome by towing undulating bodies or multi-sensor chains. Traditional current measurements at single fixed
positions have been widely replaced by acoustic profiling current meters.
An optimal ocean observing and prediction system would continuously have
access to real time data. Many measurement systems and platforms that are used
for an REA campaign, are also suitable for permanent monitoring, and it is only the
costs for maintenance and periodic replacement that prevents constant deployment.
Sensor fouling can, however, severely degrade data such as electrical conductivity,
with the consequence that the precision required for ocean mode ling is not guaranteed after weeks of exposure to a shallow littoral environment.
11.7.2 Adaptive sampling and assimilation
In an REA operation, it is not totally predictable when and from what location
updating measurements will arrive. This might put some strain on data assimilation
procedures as compared with assimilation from fixed station networks. an the
other hand, smart assimilation of data from changing locations has the potential for
detection and refinement of structures that might falI through the meshes of an
immutable observational scheme.
Free decision on the locations for the next day's measurements offers the opportunity for optimal track design. Ocean areas that are suspected or estimated to be
incorrectly represented in the model results will be covered together with areas that
have not been visited for a longer period. Adaptive sampling significantly reduces
the demands for resources and the number of platforms, which otherwise would
have to provide full area coverage with acceptable forecast degradation everywhere. Automated adaptive sampling, which is in its infancy, can be expected to
rine. Autonomous underwater vehicles (AUV) will gain more importance even for
overt surveys in future when the maximum range of small affordable units
increases to operationally meaningful distances. For time series at fixed locations,
bottom moored systems are appropriate, deployed from aircraft, ship or AUV. Following a pre-programmed sequence, they release a surface float with an antenna
that transmits the most recent data via satellite to the laboratory (Tyce et al, 1998).
Moorings of this kind are required also for non-denied observations because traditional moored instruments would store data during the full deployment period.
Data are read out only after recovery and thus not available in time for ocean prediction.
Open military environmental investigations like any civili an ocean research
activity mostly rely on ships as the main platforms. The expenses for ship time are
easily compensated by the advantage of hands-on interaction with measurement
systems. an a ship it is easier to meet the requirements for instrumentation, which
need not be as complex and sophisticated as in remotely operated systems.
For the analysis ofthe physical ocean, the direct measurement ofthose quantities
is most valuable that are represented in the fields of numeric al models, namely
temperature, salinity and water velocity. Temperature and conductivity below the
sea surface require direct probing. The limitation to one-dimensional measurements when in situ probes are lowered from a ship, is overcome by towing undulating bodies or multi-sensor chains. Traditional current measurements at single fixed
positions have been widely replaced by acoustic profiling current meters.
An optimal ocean observing and prediction system would continuously have
access to real time data. Many measurement systems and platforms that are used
for an REA campaign, are also suitable for permanent monitoring, and it is only the
costs for maintenance and periodic replacement that prevents constant deployment.
Sensor fouling can, however, severely degrade data such as electrical conductivity,
with the consequence that the precision required for ocean mode ling is not guaranteed after weeks of exposure to a shallow littoral environment.
11.7.2 Adaptive sampling and assimilation
In an REA operation, it is not totally predictable when and from what location
updating measurements will arrive. This might put some strain on data assimilation
procedures as compared with assimilation from fixed station networks. an the
other hand, smart assimilation of data from changing locations has the potential for
detection and refinement of structures that might falI through the meshes of an
immutable observational scheme.
Free decision on the locations for the next day's measurements offers the opportunity for optimal track design. Ocean areas that are suspected or estimated to be
incorrectly represented in the model results will be covered together with areas that
have not been visited for a longer period. Adaptive sampling significantly reduces
the demands for resources and the number of platforms, which otherwise would
have to provide full area coverage with acceptable forecast degradation everywhere. Automated adaptive sampling, which is in its infancy, can be expected to